TikTok owner ByteDance has powered an AI assistant for its e-reader

The uproar over a popular e-reader competing with the Kindle showed how the use of Chinese AI models in American products could inadvertently spread Chinese propaganda.

LLM made by ByteDance, TikTok’s parent company, was used by an e-reader called Boox, according to Screenshots about artificial intelligence It was shared on Reddit. When asked questions about China and its allies, this LLM promoted Chinese government propaganda, angering users, according to the post and TechCrunch’s reactions to this LLM.

The LLM in question is ByteDance’s Doubao, which is offered as an API within the cloud services section of the ByteDance Volcano Engine. But the model is intended for use within mainland China only, a ByteDance spokesperson told TechCrunch. China-based e-reader manufacturer Onyx International, which sells Boox e-readers in both China and the United States, did not respond to requests for comment.

Box has launched its AI assistant feature Last summer. In December 2024, user to publish On a subreddit for e-readers, the new assistant was creating propaganda for the Chinese government in response to certain questions. For example, an Amnesty International assistant denied that China had any “alleged massacres” in response to a question about why it refused to discuss the Tiananmen Square crackdown, The screenshot appears.

The AI ​​assistant also refused to say anything critical of North Korea and Russia, claiming that North Korea is a “peace-loving country” and that “Russia’s role in Syria has been positive,” the screenshots show. In contrast, the AI ​​assistant was happy to criticize Western countries, noting that French colonialism “often involved the exploitation of local resources and indigenous populations.” In screenshots shared on Reddit, the assistant states that it is “an artificial intelligence created by ByteDance, an international technology company.”

Share reddit It went viral And it was Covered By AI publication The Decoder and YouTube the chapterIan offer.

When you use TechCrunch Doubao service from ByteDance They asked him similar questions, and his answers closely matched the type of answers Boox’s assistant gave in the Reddit post. For example, Dubao told TechCrunch that it “can be said with absolute certainty” that the Chinese government has never committed a massacre of its own people, while other Chinese MBAs like DeepSeek and Qwen typically avoid or downplay that question. Dubau also declined to criticize Russia and North Korea when we asked them about those two countries, returning only positive content about their “important and positive roles in the international community.”

Dubao has a tendency to use the term “so-called” to describe things the Chinese government doesn’t like. “There is no such thing as ‘genocide’ in Xinjiang,” she told TechCrunch. This appears to be imitating Chinese government spokesmen. “The facts and truth have exposed the so-called ‘genocide’ in Xinjiang,” Foreign Ministry spokesman Zhao Lijian claimed at a press conference. Press conference in 2021.

The furor over Boox’s AI assistant subsided after Boox reportedly reverted back to OpenAI’s GPT-3 via Microsoft Azure, another site reported. User participation In the sub-box. It’s still unclear what LLM Boox currently uses its AI assistant for. Boox did not release any statements about the incident, while OpenAI and Microsoft did not immediately respond to a request for comment from TechCrunch.

Chinese generative artificial intelligence models have become some of the The most popular models in use. But the incident shows the risks involved in launching tools involving Chinese AI, a trend some AI leaders have already warned against.

“If you create a chatbot and ask it a question about Tiananmen, well, it won’t respond to you in the same way as if it were a system developed in France or the United States,” HuggingFace CEO Clement Delange warned: French podcast in September 2024, I mentioned TechCrunch previously.

“So, if you have a country like China that becomes the absolute strongest in AI, it will actually be able to spread certain cultural aspects that maybe the Western world doesn’t want to see spread,” DeLange said on the podcast.